Vehicular manslaughter charges filed in Los Angeles earlier this year mark the first felony prosecution in the US of a fatal car crash involving a driver-assist system.
In late 2019, Kevin George Aziz Riad’s car sped off a California freeway, ran a red light, and crashed into another car, killing the two people inside. Riad’s car, a Tesla Model S, was on autopilot.
The Los Angeles County prosecutors filed two charges against Riad, now 27. The case is also the first criminal prosecution of a crash involving Tesla’s autopilot function, which is found on over 750,000 cars in the US. Meanwhile, the crash victims’ family is pursuing civil suits against both Riad and Tesla.
Tesla is careful to distinguish between its autopilot function and a driverless car, comparing its driver-assist system to the technology airplane pilots use when conditions are clear.
“Tesla autopilot relieves drivers of the most tedious and potentially dangerous aspects of road travel,” states Tesla online. “We’re building autopilot to give you more confidence behind the wheel, increase your safety on the road, and make highway driving more enjoyable.… The driver is still responsible for, and ultimately in control of, the car.”
The electric vehicle manufacturer clearly places the onus of safety on the driver, but research suggests that humans are susceptible to automation bias, an over-reliance on automated aids and decision support systems.
Now it’s up to the courts to decide who is culpable when the use of those systems results in fatal errors. Currently, Riad is out on bail and pleading not guilty to manslaughter charges.
Here, Mark Geistfeld, professor of civil litigation at New York University, and the author of the new paper in the California Law Review, talks about the significance of the criminal charges and what they might mean for the future of consumer trust in new tech:
Can you shed some light on the legal precedent the criminal prosecution of Kevin George Aziz Riad sets? What message does it send to consumers and manufacturers of similar technology?
First, the criminal charges are surprising, based on what we know—the criminal charging documents, as usual, provide no details. Typically, if you weren’t paying attention, ran a red light and hit somebody—as tragic as it is—you wouldn’t get a criminal charge out of that behavior in the vast majority of cases. You really don’t see many criminal prosecutions for motor vehicle crashes outside of drunk-driving cases.
If the driver was found guilty of manslaughter, this case could really be the most disruptive, the most novel, the most groundbreaking precedent. It’s a strong departure from the past, if in fact the criminal prosecution is simply based on his relying on autopilot when he should have taken over. If that’s what is going on, you might see a lot more criminal prosecutions moving forward than we do today.
Tort liability, or civil charges, by contrast, is very commonplace. That’s when the defendant would pay damages for injuries caused. The majority of tort suits in state courts across the country are from motor vehicle crashes in which one driver is alleged to have negligently caused the crash, which clearly occurred in this case because the driver went through a red light.
If this case somehow signals that criminal liability is more possible simply by relying on the technology, then that could become a profound shift in the nature of legal liabilities moving forward.
What obligation does an advanced tech company such as Tesla have in informing drivers, whether directly or through advertising and marketing messages, that they are liable for all damages, regardless of whether the car is on autopilot?
They clearly have an obligation to warn the person sitting in the driver’s seat to take over the vehicle—that it’s not capable of doing everything on its own. You see that warning in Tesla vehicles, and almost all vehicles have that type of warning. For example, when you use a map function while driving, many cars will offer a warning: “This will distract you, pay attention to the road.”
Manufacturers also have an obligation to keep in mind the sense of complacency that comes with driving technology while designing the car. Tesla or any other manufacturers can’t just say, “Hey, pay attention, that’s your responsibility.”
They actually have to try to put something into the design to make sure that drivers are staying attentive. So different manufacturers are taking different approaches to this problem—some cars will pull over if your hands are not on the steering wheel, and other cars have cameras that will start beeping if you’re not paying attention.
Under current law, if the driver gets in a crash and there was an adequate warning, and the design itself is adequate enough to keep the driver attentive, the car manufacturer is not going to be liable. But there’s one possible exception here: there is a formulation of the liability rule that is pretty widely adopted across the country, including in California, where this case will take place. Under this rule, the inquiry is based on what consumers expect the manufacturer to do. And consumer expectations can be strongly influenced by marketing and advertising and so on.
For example, if Tesla were to advertise that autopilot never gets in a crash, and then a consumer does get in a crash, Tesla would be liable for having frustrated those expectations.
In this case, the driver was charged based on the idea that he was over-reliant on his car’s autopilot. What does this say about our basic assumptions about whether humans or tech are more trustworthy?
There’s an important distinction between overreliance and complacency. I think complacency is just a natural human reaction to the lack of stimulus—in this case, the lack of responsibility for executing all of the driving tasks. You can get bored and lulled into a sense of complacency, but I don’t think that behavior is being overly reliant on technology.
The idea of overreliance comes into play with the potential nature of the wrongdoing here. Maybe the driver in this case will defend himself by saying he reasonably thought the car had everything under control, was fully capable of solving this problem, and so he didn’t have to worry about reacting if things turned out otherwise.
Now at that point, he would be placing his faith in the technology instead of in his own ability to stop the vehicle and get out of the problem in a safe way. If there is blind faith in the technology rather than in taking over when you could have done so, and if you are liable as a consequence, that becomes a very profound, interesting kind of message that the law is sending.
Do you think this shift in liability will hurt business for companies like Tesla?
The big issue that autonomous vehicle manufacturers like Tesla face right now is gaining consumer trust when they’re introducing a new technology to the market. The need for trust in the early stages of these products is massively important. And all the manufacturers are worried about that problem because they know that if there are some horrific crashes, consumers are going to lose trust in the product.
Ultimately the technology will end up taking over; it’s just a question of whether it’s sooner rather than later. And time is money in this context—so if you just get slower adoption because consumers are very concerned about the safety performance of the technology, that’s going to hurt the industry. They obviously want to avoid that outcome. This technology is still going to take over—it’s just a question of how long it takes for that to happen. There are just so many advantages to using autonomous vehicles, including in the safety dimension.
Of its autopilot and full self-driving capability, Tesla says: “While these features are designed to become more capable over time, the currently enabled features do not make the vehicle autonomous.” What liability issues do you foresee if/when these vehicles do become autonomous?
It’s a complicated question, and that is the issue that everybody is interested in. Once these vehicles become fully autonomous, then there’s just the car. The human in the car isn’t even an element in the situation.
So the big question is: once those vehicles crash, who pays? You’d think the manufacturer would be liable—and that’s going to increase the cost of these vehicles and make them a lot harder to distribute. There are a lot of people who think that in the event of a crash, the manufacturer should be liable all of the time. I am strongly skeptical about that conclusion, because I think it’s a much closer call than most people make it out to be.
Ultimately, these issues depend on how federal regulators like the National Highway Traffic Safety Administration regulate the vehicle. They will have to set a safety performance standard which the manufacturer has to satisfy before it can commercially distribute the product as fully autonomous.
The question is where the regulators set that standard at, and I don’t think it’s easy to get right. At that point there will be a good debate to be had: Did they get it right or not? We’re still a few years out. I think we’ll all be having these conversations in 2025.